JP7453788B2 - How to control pink color formation during antibody production - Google Patents

Description

CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Application No. 62/503,615, filed May 9, 2017, which is incorporated herein in its entirety.

Monoclonal antibody (mAb) therapeutics are becoming more popular in the treatment of multiple human diseases. As macromolecules, mAbs have some degree of heterogeneity due to a wide range of post-translational modifications. Successful control of the manufacturing process is important to ensure product quality and safety and lot-to-lot consistency of therapeutic proteins. In fact, the ICH Q6B guidelines require a product appearance sheet.

In normal antibody production, there is generally acceptable color variation (Derfus GE, et. al., MAbs 2014, 6(3):679-688, Prentice KM, et. al., MAbs 2013, 5). 6):974-981, Vijayasankaran N, et. al., Biotechnol Prog 2013, 29(5):1270-1277, Xu J, et. al., process Biochemistry 2014, 49:1 30-1394). However, color variations due to in-process impurities are a concern due to lack of process control. One of the main causes of the pink/red color of the final drug substance has been identified as the vitamin B ₁₂ -mAb complex (Derfus, 2014, Prentice, 2013). Although the nature of the interaction has not been elucidated, the attachment of vitamin _B12 to proteins has been demonstrated using multiple methods, including protein A affinity chromatography, low pH viral inactivation, various polishing chromatographs, and ultrafiltration and diafiltration. It appears to be strong enough to co-elute throughout downstream purification steps.

Antibodies are commonly produced in mammalian cells, such as Chinese hamster ovary (CHO) cells, and are secreted extracellularly into the cell culture medium. At the end of the culture step, cells are separated from the culture medium during a primary recovery step using methods such as centrifugation, depth filtration, or flocculation to clarify the collected liquid. During this process, cells are often subjected to various stresses including mechanical shear, exposure to low dissolved oxygen (DO) environments, or temperature and pH shifts. These stresses cause cell damage and release of unwanted intracellular components into the clarified fluid. These cytosolic components such as lipids and enzymes can affect product quality and must be carefully monitored or removed. For example, release of intracellular reducing agents can lead to disulfide bond reduction of antibodies. (Mun M., et.al. Biotechnol Bioeng 2015;112:734-742; Ruaudel J, et.al. BMC Proceedings 2015;9(Suppl 9):P24; Mullan B, et.al. BM C Proc. 2011; 5(Suppl 8):P110; Trexler-Schmidt M, et.al. Biotechnol Bioeng. 2010;106:452-461; Koterba, KL, et.al. J Biotechnol. 2012;157:26 1-267;Handlogten MW, et.al. Biotechnol Bioeng. 2017;114:1469-1477; Hutchinson N, et.al. Biotechnol Bioeng. 2006;95:483-491).

During the manufacturing process, significant antibody reduction was observed after the collection procedure and/or Protein A chromatography. Multiple process parameters may affect the extent of antibody reduction. For example, maintaining high levels of dissolved oxygen (DO) during collection is essential to keep antibody molecules intact (Mun M. 2015; Trexler-Schmidt M, 2010). Mechanical shear forces that cause cell lysis and leakage of cell components into the collection fluid also play a major role in reduction (Kao YH, et. al. Biotechnol Bioeng. 2010; 107:622-632; Hutterer KM, et. al. MAbs. 2013;5:608-613). Collection retention time (Chung WK, et.al. Biotechnol Bioeng. 2017; 114:1264-12741), media components (copper ions, cysteine/cysteine, etc.) and pH and temperature (Trexler-Schmidt M, 2010; Chung, 2017) Other process parameters also influence the extent of disulfide reduction.

To ensure the quality of antibody products, manufacturing controls during manufacturing are necessary to control low molecular weight (LMW) species formed from reduction of antibody disulfide bonds. As a result, several strategies have been proposed in recent years to control disulfide reduction. Chemical inhibitors have been tested to prevent antibody reduction, including copper sulfate (Chaderjian WB, et. al. Biotechnol Prog. 2005;21:550-553), ethylenediaminetetraacetic acid (EDTA), and thioredoxin inhibitors (U.S. Patent No. 8,574,869) prior to collection with anti-reducing agents such as cysteine, methyl blue (WO2015/085003) and coenzyme Q analogs (Li WW, et.al. J Am Chem Soc. 2005;127:6140-614). and post-collection processing. Although the knowledge and methods surrounding mitigation strategies have increased over the years, the implementation of these methods in production involves the introduction of chemical byproducts that must be removed by chromatographic steps, increased processing times, and off-target modifications or This is not without its challenges, including the risk of damaging antibody products.

The present invention includes any of the following;
1) prevent reduction of the antibody during production and collection, or 2) inhibit the conversion of cyanocobalamin (CN-Cbl) to hydroxocobalamin (HO-Cbl) during production and collection, or 3) prevent production and A method of preventing the formation of a pink color during antibody production both by preventing reduction of the antibody during collection and by inhibiting the conversion of cyanocobalamin to hydroxocobalamin.

In one embodiment of the invention, CN-Cbl conversion to HO-Cbl prevents medium exposure to visible light during one or more of the following steps selected from medium preparation, medium storage, antibody production, and antibody collection. inhibited by

In one embodiment of the invention, CN-Cbl is exposed to UVA light only during one or more times selected from media preparation, media storage, antibody production, and antibody collection.

In one embodiment of the invention, CN-Cbl is exposed to only red light (>60 On m) during media preparation, media storage, and optionally during antibody production and antibody collection.

In one embodiment of the invention, reduction of antibody disulfide bonds is prevented by adding hydrogen peroxide to the clarified bulk at the time of collection.

Another embodiment of the invention is to 1) prevent the reduction of antibody disulfide bonds during manufacturing and collection, or 2) HO during cell culture media preparation, media storage, cell culture steps, or antibody collection. - A method of inhibiting the binding of vitamin _B12 to an antibody during manufacturing, comprising inhibiting the conversion of CN-Cbl to Cbl.

Another embodiment of the invention includes: 1) preparing and storing a cell culture medium under red light (>600 nm) conditions; 2) culturing cells producing antibodies of interest under red light conditions. 3) collecting antibodies from a cell culture under red light conditions; and 4) adding hydrogen peroxide to the collection solution.

Another embodiment of the invention includes: 1) preparing and storing a cell culture medium under red light (>600 nm) conditions; 2) culturing cells producing antibodies of interest under red light conditions. and 3) harvesting the antibody of interest from a cell culture under red light conditions.

Another embodiment of the invention comprises: 1) culturing cells producing the antibody of interest; 2) collecting the antibody of interest from the cell culture; and 3) adding hydrogen peroxide to the collection solution. A method for the production of antibodies, including.

Figures 1A-1D show the spectral absorption of two lots of vitamin B ₁₂ and mAb B. 1A is the absorption spectrum of two lots of vitamin B ₁₂ and mAb B. Vitamin B ₁₂ (cyanocobalamin) (blue line) at 5 mg/L in PBS from 250 to 800 nm, pink lot (red line) and colorless lot (purple line). 1B is a rescaled plot from (1A) to show details. 1D is a bottle of Protein A eluate for mAb B with a visible pink color. Figures 2A-2D show that the attachment of vitamin _B12 to antibodies is a secondary reaction. (2A) Two lots of colorless mAb A, one with no detected reduced antibody molecules (a "good" lot) and another with a majority of reduced antibodies (a "bad" lot), were treated with CN-Cbl or Mixed with either HO-Cbl. After incubation, excess free vitamin B12 was removed. The samples are from left to right: good lot and CN-Cbl; bad lot and CN-Cbl; good lot and HO-Cbl; bad lot and HO-Cbl; and the bad lot was first treated with hydrogen peroxide; Mixed with HO-Cbl. (2B) Hydrogen peroxide can oxidize reduced antibodies and reform antibody monomers, as shown by CE SDS. (2C) Quantification of mAb monomer and various low molecular weight species in hydrogen peroxide treated and untreated mAbs shown in 2B. (2D) Non-reduced CE-SDS analysis of mAb A batch 2 material before (blue line) and after (black line) vitamin _B12 attachment. The individual peak patterns changed before and after incubation with vitamin _B12 , probably due to oxidation of free sulfhydryl groups in air and reformation of disulfide bonds. Figures 3A-3C show mass spectral data for vitamin _B12 conjugation identification. (3A) MS/MS spectra of m/z = 664.79 from HO-Cbl and m/z = 678.29 from CN-Cbl. (3B) MS/MS spectra of peptide-cobalamin complexes L19-cobalamin and L18-19-cobalamin. (3C) Diagram of IgG4 antibody disulfide bonds. Cysteine residues that bind vitamin _B12 are shown in blue, and pairs of cysteine residues that form disulfide bonds are circled in red. 4A-4E show the light transmission spectra of the color filters used in Example 1. (4A) Chromatogram for separation on RP-UPLC and detection at 360 nm (4B) CN-Cbl and HO-Cbl in standards; (4C) fresh culture medium for mAb A; (4D) exposure to red light. and (4E) medium exposed to green light. In (4D and 4E), the black, blue and red lines indicated the exposure energy of 0.3, 0.6 and 1.2 million lux hours, respectively. Figures 5A-5C show free sulfhydryl levels and low molecular weight (LMW) in CB. (5A) Correlation of free thiol levels in mAb 1 CB samples and the presence of LMW in the corresponding Protein A purified samples (PAVIB). PAVIB was analyzed by NR_Caliper for LMW species. Those samples with LMW≧5% are labeled as “Yes”, otherwise “No”. Analysis was performed using JMP software. (5B) Measurement of free thiol levels in CB under different collection treatments. Clarified harvest bulk (CB) from a laboratory scale bioreactor (5-L) containing mAb 1 was treated by lowering the pH to 4.8 in the absence or presence of dextran sulfate. CB samples were incubated for 1 hour at room temperature before the samples were depth filtered. After depth filtration, an aliquot of each sample was incubated for 75 minutes at 37° C. with or without the addition of 1 mM NAPDH before free sulfhydryl content was measured. Production lots with a high percentage of LMW and representative lots (regular lots) are also included for comparison. (5C) Percentage of intact antibodies for samples in (5B). Intact antibodies were measured by NR_Caliper after Protein A purification. Figures 6A-6F show redox indicator DCPIP treatment and observation of color change as a predictive marker for LMW production. (6A) Structure and reaction of DCPIP. (6B) On the left, mAb2 purified DS samples with >95% LMW (3.1 sulfhydryl groups per antibody protein) showed colorlessness when treated with DCPIP. The mAB2 sample on the right with more than 98% intact antibody (0.4 sulfhydryls per antibody protein) showed a purple color (at pH 5.5) in the presence of DCPIP. (6C) DCPIP color change in the CB of mAb 2 from different culture conditions. DCPIP in the two tubes on the left was completely colorless, indicating that a high reduction potential was present in the CB; DCPIP in the two tubes in the middle had a blue color, indicating no reducing effect; and on the right. The DCPIP in the tube was partially reduced. (6D) Free thiol concentration as a function of mAb 2 percentage of cell lysis. 100% cell lysates of mAb 2 were mixed with normal CB, cell lysate dilution curves were generated, and free thiol concentrations were measured and reported as the average of duplicate samples. The linear regression equation and R-square were calculated using Excel. (6E) Study design to test the order of color change and free thiol content of DCPIP. (6F) Image of DCPIP test results for the studies in Table 7. The samples in the top panel were aerated before and during primary collection. The bottom sample was subjected to nitrogen gas flushing before and during primary collection. The order of the test tubes from left to right was the same as the order in Table 7 from top to bottom. Figures 7A-7K demonstrate the use of hydrogen peroxide to prevent disulfide bond reduction. mAb 2 Laboratory-scale generated CB was dispensed into containers of hydrogen peroxide at various concentrations. Airless conditions within the container were created by nitrogen flushing and kept at room temperature for 1 day. The resulting samples were directly analyzed with non-reduced (7A, 7C, 7E, 7G, 7I) and reduced (7B, 7D, 7F, 7H, 7J) Caliper without Protein A purification. (7A and 7B) CB without addition of hydrogen peroxide was exposed to air as a control. CB with 0mM hydrogen peroxide (7C and 7D), CB with 0.33mM hydrogen peroxide (7E and 7F), CB with 1mM hydrogen peroxide (7G and 7H), and CB with 3mM hydrogen peroxide (7I and 7J) was held in an airless state. (7K) Summary of non-reducing Caliper results for mAb fragmentation. Because there is an excess of light chain of this mAb secreted from the host cell. The amount of light chain is not included as fragments. Abbreviations used for LMW species: LC, light chain; HC, heavy chain; HL, half antibody of one light chain and one heavy chain; HHL, partial antibody of one light chain and two heavy chains. Figures 8A-8F show evaluation of using hydrogen peroxide to prevent disulfide bond reduction in the worst case. The mAb 2CB tested had 100% cell lysis as described in Example 2. The holding conditions with and without air were the same as in FIG. (8A) CB samples kept without hydrogen peroxide treatment and with air serve as controls. (8B) CB sample without hydrogen peroxide treatment and held in airless condition. (8C) 5mM hydrogen peroxide was added to the CB samples prior to holding in airless conditions. (8C) 10 mM hydrogen peroxide was added to the CB samples before holding in airless conditions. The resulting CB samples (8A to 8D) were analyzed by direct non-reducing Caliper without Protein A purification. (8E) Summary of non-reducing Caliper results of mAb fragmentation from unpurified CB. (8F) Summary of intact antibody purity of samples 8A to 8D purified by Protein A chromatography and analyzed by non-reducing Caliper. Figures 9A-9D demonstrate the use of alternative peroxides to prevent antibody disulfide bond reduction. Sodium percarbonate and sodium perborate were used in the worst case of mAb 2 CB with 100% cell lysis. (9A) CB sample held with air and no peroxide. (9B) CB sample kept airless and without peroxide. (9C) Representative results for 5mM sodium percarbonate treated CB samples. Sodium percarbonate was added prior to the airless hold. Results from unreduced Caliper of crude CB (9A to 9C). (9D) Overview of sodium percarbonate and sodium perborate treatments. Results from unreduced Caliper of unpurified CB.

DETAILED DESCRIPTION OF THE INVENTION Process control in biologics manufacturing is important to ensure product quality and safety and lot-to-lot consistency of therapeutic proteins. Understanding the mechanisms of color production during manufacturing is important for developing control strategies. We discovered that the formation of a pink color during antibody production is a secondary reaction that depends on the concentration of both the free sulfhydryl groups of the reduced antibody and hydroxocobalamin, the active form of vitamin _B12 . Both reactants are necessary; either alone is not sufficient to produce the pink color. The present invention: 1) prevents reduction of antibodies during manufacturing and collection; or 2) inhibits the conversion of cyanocobalamin (CN-Cbl) to hydroxocobalamin (HO-Cbl) during manufacturing and collection. This is a method of preventing the formation of a pink color during antibody production by either of the following: Alternatively, the present invention provides a method for producing a pink color during antibody production by 1) preventing reduction of the antibody during production and collection, and 2) inhibiting the conversion of cyanocobalamin to hydroxocobalamin during production and collection. This is a method to prevent the generation of

We found that the active form of vitamin _B12 is located at heavy chain 134 (HC134), light chain 214 (LC214), heavy chain 321 (HC321), heavy chain 367 (HC367), and heavy chain 425 (HC425). It was discovered that the free sulfhydryl group of cysteine is attached to the free sulfhydryl group of cysteine. The five cysteine residues are distributed in both the Fab and Fc regions. There are two pairs of disulfide bonds between these five cysteine residues, LC214 and HC134 and HC367 and HC425 (Figure 3c). When these particular disulfide bonds are reduced during manufacture and collection, both cysteine residues in the disulfide pair have equal accessibility for hydroxocobalamin attachment.

As used herein, "cyanocobalamin" and "CN-Cbl" are used interchangeably and refer to the vitamin B ₁₂ (VB12) component of cell culture media. Similarly, "hydroxocobalamin" and "HO-Cbl" are used interchangeably to make the active form of vitamin _B12 (VB12) available in the cell culture medium to bind reduced antibodies during manufacture and collection. Point.

Antibodies are usually produced in mammalian cells, such as Chinese hamster ovary (CHO) cells, and secreted extracellularly into the cell culture medium. At the end of the cell culture process, the cells are separated from the culture medium during a primary recovery step using methods such as centrifugation, depth filtration, or flocculation to clarify the collected liquid.

As used herein, "clarified bulk" and "CB" are used interchangeably and refer to primary processes such as centrifugation, depth filtration, or flocculation used to clarify cell culture media. Refers to the solution recovered after the recovery step.

As used herein, "cell culture media preparation,""cell culture media storage,""cell culture process,""antibodyproduction,""antibodyproduction," and "antibody collection,""collection ” refers to the many steps necessary to produce antibodies. Each step may occur at a different location within a manufacturing facility.

Using red light in manufacturing environments as a pink color control strategy

Since both hydroxocobalamin and reduced antibody are required for the production of a pink product, control of either one or both factors will prevent pink production.

Light-induced vitamin _B12 conversion can occur during media preparation, media storage, and cell culture steps. The use of transparent disposable bioprocessing bags during media preparation, storage, and cell culture in glass reactors increases the risk of vitamin _B12 light exposure and conversion. Common control strategies such as using appropriate containers (e.g., stainless steel containers), covering with additional layer materials, and avoiding unnecessary light exposure can provide protection from light. However, implementing these strategies is not without its challenges and often involves additional costs.

Example 1 shows that red light does not induce CN-Cbl conversion to HO-Cbl. Reduced mAb A was spiked into culture medium exposed to white light, or red or green light, or kept in the dark. After overnight incubation, mAbs were purified on a protein A column and bound vitamin _B12 was measured in an ELISA assay. As shown in Table 7, the order of vitamin B ₁₂ conjugation rates was white light > green light > red light, and dark control.

In one embodiment of the invention, cyanocobalamin conversion to hydroxocobalamin is performed using white light (wavelength >600 nm) in areas where cell culture media are prepared and stored, and optionally in cell culture production and collection areas. inhibited, reduced, or prevented by replacing it with

In another embodiment of the invention, cyanocobalamin conversion to hydroxocobalamin is performed by replacing white light bulbs with red LED light bulbs (wavelength > 600 nm) is inhibited, reduced or prevented.

In another embodiment of the invention, cyanocobalamin conversion to hydroxocobalamin is carried out on a red plastic sheet or Inhibited, reduced or prevented by placing a red tube.

In another embodiment of the invention, the white light in the area where the cell culture media is prepared and stored, and also in the cell culture production and collection area, is replaced by a white light bulb or a red LED light bulb (wavelength >600 nm). or by placing a red plastic sheet or red tube over a white light bulb.

In another embodiment of the invention, the white light in the area where the cell culture media is prepared and stored and also in the cell culture collection area is replaced by a white light bulb with a red LED light bulb (wavelength >600 nm). Red light can be replaced by replacement or by placing a red plastic sheet or red tube over a white light bulb.

In another embodiment of the invention, the white light in the area where the cell culture media is prepared and stored and also in the cell culture production area is replaced by a white light bulb with a red LED light bulb (wavelength >600 nm). Red light can be replaced by replacement or by placing a red plastic sheet or red tube over a white light bulb.

Prevention of antibody reduction during protein collection and primary recovery as a pink color control strategy.

Since both reduced antibody and hydroxocobalamin are required for the production of a pink product, control of either one or both factors will prevent pink production.

Disulfide bond reduction during mAb production is an enzymatic redox reaction and is the root cause of low molecular weight species (LMW) formation during the antibody production process. Although the causes of LMW generation are not well understood, the appearance of LMW is attributed to the combined effects of cell culture conditions such as temperature, dissolved oxygen (DO), pH, viability, cell density, and lysis rate, which create a reducing environment for antibodies. It is considered. This is further exacerbated by clarified bulk (CB) holding conditions (eg, temperature, time, aeration) that hold antibodies in a reducing environment. Detection of LMW is usually visualized in the clarification bulk or protein A eluate pool step.

Example 2 shows that free thiol levels below 100mM have a low risk of LMW formation; free thiol levels between 100 and 200mM are a warning sign; and free thiol levels above 200mM have a low risk of LMW formation. Indicates that it is high. Many mitigation strategies to prevent antibody reduction have been proposed to inhibit enzymes involved in redox reactions or to oxidize and deplete important enzyme cofactors such as NADPH (Mun M 2015, Trexler-Schmidt M 2010). Air sparging has been shown to be a powerful method to prevent antibody reduction. However, total free thiol levels vary from micromolar levels during production to millimolar levels after holding the CB without air (Table 7). Figure 8 shows that for the worst clarified bulk conditions, exposure to air could not completely suppress antibody reduction.

We discovered that adding peroxide to the clarified bulk during collection prevented the reduction of antibody disulfide bonds and the generation of LMW species. Advantages of using peroxide include 1) peroxide can be mixed with water in any ratio, and 2) removal of excess peroxide is not burdensome to downstream purification. However, when a 10 mM solution of hydrogen peroxide (H ₂ O ₂ ) completely decomposes into water and oxygen, 5 mM of oxygen can theoretically be produced, but this process is slow in the absence of a catalyst.

Example 2 shows that adding hydrogen peroxide or alternative inorganic or organic peroxides such as sodium percarbonate or sodium perborate to the clarified bulk can effectively prevent the reduction of antibody disulfide bonds. show.

In one embodiment of the invention, antibody disulfide bond reduction is prevented by adding hydrogen peroxide to the clarified bulk at the time of collection.

In one embodiment of the invention, antibody disulfide bond reduction is prevented by adding an inorganic or organic peroxide, such as sodium percarbonate or sodium perborate, to the clarified bulk at the time of collection.

In another embodiment of the invention, antibody disulfide bond reduction is prevented by maintaining the hydrogen peroxide concentration of the clarified bulk at ≦10 mM during collection.

In another embodiment of the invention, antibody disulfide bond reduction is prevented by maintaining the hydrogen peroxide concentration of the clarified bulk between 3 and 10 mM during collection.

Example 1
Vitamin B12 Binding with mAbs Protein Stocks Two different therapeutic IgG4 molecules (mAbs A and B) were used in this study. Both of them had one point mutation from serine to proline in the original IgG4 hinge region motif CPSC, resembling the IgG1 interchain disulfide bond structure. (Liu H et. al.: MAbs 2012, 4(1): 17-23; Aalberse RC et. al. Immunology 2002, 105(1): 9-19). IgG4 antibodies were produced in CHO cells and affinity purified using protein A chromatography, additional polishing chromatography and final UF/DF filtration into histidine-based buffer. Protein concentration of drug substance (DS) was determined by absorbance at 280 nm.

Vitamin _B12 ELISA.
A commercially available VB12 assay kit (Monobind Inc., Lake Forest, Calif.) was used in conjunction with the manufacturer's recommended procedures to determine the amount of vitamin B ₁₂ (VB12) bound to the purified protein.

Vitamin _B12 conjugation to drug substance in vitro.
mAb A or mAb B solution (50 mg/ml) in histidine-based buffer at a molar ratio of 4:1 to 2:1 (protein:VB12) with CN-Cbl (Sigma-Aldrich, V2876, stock solution 1 mg/ml water) or Mixed with either HO-Cbl (Sigma-Aldrich, V5323, stock solution 1 mg/ml water). The reaction was carried out at room temperature in the dark for 1 day. Buffer exchange with phosphate saline buffer (PBS, Sigma-Aldrich, D5652) was performed through a 50 kDa cutoff Amicon centrifugal filter until no color was visible in the flow-through. The buffer exchange reaction solution was subjected to peptide mapping analysis.

Hydrogen peroxide oxidation of reduced antibodies.
One lot of mAb A had disulfide bonds that were partially reduced during manufacturing. The drug substance was mixed with 0.1% hydrogen peroxide (30 mM) for 4 hours at room temperature and excess hydrogen peroxide was removed by buffer exchange as described above.

Light exposure studies.
Freshly prepared medium for mAb A was used in light exposure studies. CN-Cbl was added to a final concentration of 10 mg/L (from a 1 mg/ml stock solution).

A 13W white light compact fluorescent lamp (CFL) was used as the visible light source. A 13W commercially available black light bulb centered at 360 nm was used as the UV light source. The lamp was installed in a fume hood with an air flow rate of 200 cubic feet per minute to prevent heating of the solution. The surface of the bench was dark to avoid light reflections. The sample was placed approximately 12-13 inches from the lamp (average intensity was 672 lux at the surface of the tube) to simulate typical room lighting conditions (500-1,000 lux). Light intensity was measured with a photometer (Sper Scientific UVA/B photometer Model 850009 and Extech HD400 for UVA and visible light, respectively). Color filters were purchased from Arbor Scientific (Kit 33-0190). The transmission spectrum of the filter is shown in FIG. Polycarbonate plastic sheets (2.4 mm thick) were purchased from a hardware store and used to filter out UVA light.

All medium samples were exposed to light sources at 0.3, 0.6, and 1.2 million lux hours (intensity x exposure time) for visible light and 71 and 212 watts per square meter for UVA light. sampled.

Free sulfhydryl group assay.
The method was modified slightly from the literature (Arner ES et.al. Methods Enzymol 1999, 300:226-239) using Ellman's reagent, 3,3'-dithio-bis(6-nitrobenzoic acid), 3-carboxylic -4-nitophenyl disulfide; 2,2'-dinitro-5,5'-dithiobenzoic acid (DTNB) is used. Briefly, 10 μL of drug substance sample was mixed with 15 μL of TE buffer (50 mM Tris-HCl, 20 mM EDTA, pH 7.6) in a 96-well plate. DTNB solution (5 mg/ml in ethanol, Sigma-Aldrich, D8130) was freshly mixed with 8 M guanidine-HCl in 0.2 M Tris (pH 8) in a volume ratio of 1:9 before use. Next, 100 μL of DTNB/guanidine solution was added to each well and mixed. The spectral absorption at 412 nm was measured and a molar extinction coefficient value of 13,600 M ⁻¹ cm ⁻¹ was used for calculations.

Antibody denaturation test.
For thermal denaturation, 1 mL of mAb A (neutralized with Tris after protein A affinity column purification) was heated at 80° C. for 20 min in an Eppendorf test tube. After heating and cooling a visible precipitate formed. Tubes were centrifuged at 18,000xg for 20 minutes. The supernatant was used for protein assay (OD280) and vitamin B ₁₂ ELISA assay. For SDS denaturation, 0.25 mL of drug substance was mixed with 0.75 mL of 0.3% sodium dodecyl sulfate (SDS) in phosphate saline (PBS) and left at room temperature for 1 day. Excess SDS was removed by buffer exchange with PBS through a 50 kDa cutoff Amicon centrifugal filter.

Peptide mapping analysis of drug substance-vitamin B12 reaction mixture using LC/MS/MS.
Samples containing mAb A and vitamin _B12 were digested with trypsin without prior reduction or alkylation. Digestion buffer (50mM Tris-HCl, 10mM calcium chloride, 2M urea, pH 7.6) was used to prepare the buffer exchange reaction mixture at 1mg/wt before adding trypsin at an enzyme to protein ratio of 1:25 (wt/wt). (diluted to mL). After incubation at 37°C, the digested mixture was directly subjected to LC/MS/MS analysis. Approximately 10 μg of the tryptic digest was chromatographically separated using a Waters Acquity™ UPLC before analysis on a ThemoFisher Q-EXACTIVE PLUS Orbitrap mass spectrometer. A Waters Acquity BEH CSH(TM) C18 column (1.7 μm, 2.1×100 mm) was used for the separation (at 45° C.). Peptides were eluted in water using a linear gradient of mobile phase B from 1% to 80% over 60 min (mobile phase A: 0.1% formic acid and 5mM ammonium formate; mobile phase B: 0.1% in acetonitrile). % formic acid and 5 mM ammonium formate), flow rate 0.2 mL/min. The Orbitrap mass spectrometer was operated in positive ion mode with a spray voltage of 3.0 kV. The heater temperature and capillary temperature were set at 300°C and 275°C, respectively. MS acquisition was performed at a resolution of 70,000 and the 10 most intense ions were selected at each duty cycle at a resolution of 17,500 for MS/MS analysis. The dynamic exclusion function was effective. Fragmentation spectra were obtained using 25 normalized collision energies.

Both the disulfide-reduced mAb and HO-Cbl are required to produce the pink color.
During mAb production, it is observed that a pink color usually appears at the Protein A elution step, concentrating the mAb and removing background color from the media composition. The content of vitamin _B12 was analyzed by ELISA method and the pink batch was found to have about 20 times higher vitamin _B12 than the colorless batch (Table 1, batch 1 vs. 2).

Table 1. Vitamin _B12 (VB12) attachment and free thiol of mAb B

Based on the molecular weight and concentration of vitamin _B12 and mAb, the colored material had approximately 1 vitamin _B12 molecule in 47 mAb molecules. Vitamin _B12 has three major absorption peaks around 360, 420, and 550 nm. Spectral scans showed that the pink batch had higher absorption above 320 nm than the colorless batch (Figures 1A and 1B, red vs. purple lines).

It has been proposed that the reduction of disulfide bonds and the generation of free sulfhydryl groups is associated with the appearance of a pink color (Derfus GE et.al.. MAbs 2014, 6(3):679-688). The free sulfhydryl content of mAb B in the pink lot had higher free sulfhydryl levels than the colorless batch (Table 1). On the other hand, some batches with high free sulfhydryl content (N=4 for mAb A, N=2 for mAb B) had no pink color in the drug substance. For example, batch 2 of mAb A had higher levels of free sulfhydryl groups than batch 1, but both batches were colorless and had similar levels of vitamin _B12 (Table 2). . These results indicate that free sulfhydryl groups are necessary but alone are not sufficient to produce a pink product.
Table 2. Vitamin _B12 attachment and free thiol of mAb A

Vitamin _B12 is an essential cofactor for mammalian cell culture. CN-Cbl, the most stable form of vitamin _B12 , is the only source used in cell culture processes. It has been reported that light can convert CN-Cbl to HO-Cbl, which is highly reactive towards mAbs (Prentice KM et.al.: MAbs 2013, 5(6):974-981). Low (batch 1) or high (batch 2) levels of free sulfhydryl content to demonstrate the need for both free thiols from proteins and reactive HO-Cbl to produce pink color in the drug substance. A series of tests were set up by incubating mab A with CN-Cbl or HO-Cbl. Vitamin _B12 isoforms and mAbs were mixed and incubated in the dark at room temperature for 1 day. After incubation, free vitamin _B12 was removed by filtration. The amount of vitamin _B12 retained was (Table 2);
(1) HO-Cbl is much more active than CN-Cbl to form complexes with the same batch mAb;
(2) drug substances with higher levels of free sulfhydryl groups are more reactive than drug substances with lower levels of free sulfhydryl groups, and (3) the highest vitamin _B12 uptake is in the presence of HO-Cbl and at high levels. of drug substances with free sulfhydryl groups, causing a more than 100-fold increase in vitamin _B12 . Differences in color due to the vitamin _B12 content of each sample were visible to the naked eye (Figure 2a).

Free sulfhydryl groups are derived from reduced disulfide bonds in proteins. The drug substance of mAb A Batch 2 had an average of 3.1 free sulfhydryl groups per protein molecule (Table 2), and more than 95% of the antibody molecules were reduced (Figures 2b and 2c). Hydrogen peroxide (0.1% v/v) was added to the Batch 2 material, converting the majority (86.6%) of the low molecular weight species (LMW) to antibody monomers (Figures 2b and 2c). The reoxidized material had lower free sulfhydryl levels (measured after H2O2 treatment and before HO-Cbl incubation) and lower reactivity towards HO-Cbl. These results indicate that the formation of the pink color is a second-order reaction, the rate of which depends on the concentration of both reactants, free sulfhydryl groups and HO-Cbl.

Vitamin _B12 is covalently bound to the antibody.
Previous literature suggested a non-covalent nature, but no experimental evidence was provided regarding the nature of vitamin _B12 attachment (Derfus et al., 2014). The presumed non-covalent binding would require structural and charge complementarity of the antibody to vitamin _B12 , and thus disruption of the native protein structure would release the bound vitamin _B12 . Two denaturation methods were used: heating and SDS treatment. Heating at 85°C for 20 minutes denatured and precipitated the mAb B pink material. The nearly protein-free supernatant contained only about 3.5% of the total vitamin _B12 of the starting protein solution (Table 1). SDS denatures the protein by binding to the peptide backbone and keeping the denatured protein dissolved in solution. Vitamin _B12 assays showed that most, if not all, of the vitamin _B12 was still bound to the denatured protein (Table 1). A mobility shift of the colored material was detected in non-reducing capillary electrophoresis-SDS (CE-SDS). Both the initial mAb A drug substance and the colored complex of mAb A-vitamin _B12 had multiple bands in non-reducing CESDS due to disulfide reduction. In addition to the peaks overlapping the corresponding peaks in the initial material, the colored material had additional peaks shifted to higher molecular weight, with the shift being more pronounced for the lower molecular weight peaks such as the single light and heavy chains (Figure 2d). This observation could be easily explained as the addition of a vitamin _B12 molecule to the mAb. Taken together, these results demonstrate that vitamin _B12 is bound to the denatured protein and co-migrates with the protein under denaturing conditions. Therefore, the interaction of vitamin _B12 with the mAb is unlikely to be of a non-covalent nature.

Both CN-Cbl and HO-Cbl were observed in tryptic digests by LC/MS/MS analysis. The theoretical molecular weights of CN-Cbl and HO-Cbl are 1254.567 and 1345.567 Da, respectively. The main ion observed in CN-Cbl is a doubly charged ion with m/z 678.293, corresponding to CN-Cbl with two protons added. However, the most predominant ion observed in HO-Cbl is a doubly charged ion with m/z 664.789, corresponding to B12-OH with an OH ligand, releasing a coordination compound and adding a proton. The MS/MS spectra of the m/z 664.789 species from HO-Cbl and m/z 678.293 from CN-Cbl are shown in Figures 4A and 4B, respectively. The two MS/MS spectra have many common fragmentations, including m/z values of 147.09, 359.10, 456.72, 912.44, and 1124.45. m/z 997.48 and 1209.49 in the MS/MS spectrum of CN-Cbl are derived from the species m/z 912.44 and 1124.45 without loss of the CoCN group.

LC/MS/MS analysis of tryptic digests revealed a coordinated covalent bond between cobalamin and IgG through coordination with several cysteine residues. There are five cysteine residues in each light chain and eleven cysteine residues in each heavy chain. However, as listed in Table 3, only five of them (L-C214, H-C134, H-C321, H-367, H-C425) only. L-C214 is at the C-terminus of the light chain and is the only light chain cysteine residue involved in the formation of covalent bonds.

Table 3. Cysteine containing peptide-cobalamin complexes observed in tryptic digests of IgG-vitamin _B12 reaction mixtures

This is evidenced by the masses observed for the C-terminal peptide L ₁₉ -Cbl complex (1635.646 Da) and the incompletely digested peptide L _18-19 -Cbl complex (2139.894 Da) with a mass accuracy of 2 ppm. . These coordination complexes were also confirmed by MS/MS analysis. The MS/MS spectra of these two peptide-Cbl complexes are very similar and consist mainly of fragments from cobalamin. Most of the MS/MS fragments can be identified by comparison with the MS/MS spectrum of the HO-Cbl m/z 664.789 species. As shown in Figures 4A and 4B, m/z 147.09, 359.10, 486.24 are exactly the same as the MS/MS spectrum of HO-Cbl. Furthermore, the species at m/z 971.47 and 1183.48 consist of the same structure at m/z 912.44 and 1124.45 and cobalt (atomic mass 58.93 Da).

Similarly, it has been confirmed that the four cysteines on the heavy chain form coordinate bonds with cobalamin. H-C134 is the disulfide bond counterpart with L-C214 in the FAB region. The other three cysteine residues H-C321, HC-C367 and H-C425 are the three C-terminal cysteines on the heavy chain Fc region. Furthermore, HC-C367 and HC425 are disulfide bridges of a pair of native IgG4.

Photoconversion of CN-Cbl to HO-Cbl and process control of pink product. As mentioned above, the pink reaction is a secondary reaction that depends on the concentration of reduced antibody molecules and HO-Cbl. Therefore, process control strategies are required to prevent reduction of antibody molecules and/or generation of HO-Cbl. Controlling the reduction of antibody products is a very complex process, and has been extensively studied in antibody redesign (Peters SJ et.al. J Biol Chem 2012, 287(29):24525-24533; Aalberse RC et.al, 2002), culture media Composition (Trexler-Schmidt M et.al. Biotechnol Bioeng 2010, 106(3):452-461), process parameters (Mun M et.al. Biotechnol Bioeng 2015, 112(4):734-742) and collection treatment ( (Trexler-Schmidt M et. al, 2010). Controlling the amount of HO-Cbl by avoiding exposure to visible light is also an alternative strategy. However, these options may pose operational inconveniences and may be cell lineage specific.

＊エネルギー単位は、可視光につきＫｌｕｘ時間およびＵＶＡライト用のメーターにつきワット／ｓｑである。 We addressed this problem by identifying the wavelengths that cause CN-Cbl conversion and set up a control strategy by removing harmful wavelengths from the light source. Vitamin _B12 has both UVA and visible light absorption (Figure 1). Freshly prepared mAb A medium was exposed to (1) normal laboratory light provided by a white fluorescent tube at an intensity of 500-1,000 lux, or (2) enhanced visible light provided by a CFL fluorescent lamp, or (3) Exposure under enhanced UVA light centered at 360 nm. CN-Cbl and HO-Cbl were separated and quantified by RP-UPLC (Figure 4B). The results show that it was visible light rather than UVA light that primarily converted CN-Cbl to HO-Cbl (Table 4). The conversion rate was a function of the overall illumination (intensity x time) from both normal laboratory light and high intensity visible light.
Table 4. Visible light, not UVA light, is responsible for CN-Cbl conversion.

*Energy units are Klux hours for visible light and Watts/sq per meter for UVA light.

Next, we investigated the range of visible wavelengths (or colors) that can induce CN-Cbl conversion as a function of energy. Selected color filters were used to generate light with specified wavelength cutoffs to isolate specific regions of the light spectrum (Figure 4A). The extent of CN-Cbl conversion as a function of exposure time to different colors of light, as shown in Table 5. Although polycarbonate glass can block most UVA light, it allows visible light to pass through, so it could not prevent CN-Cbl conversion. Exposure to red light with wavelengths above 600 nm had the slowest rate of CN-Cbl conversion (Fig. 4d). In contrast, both green light (Fig. 4e) and blue light have much higher CN-Cbl conversion rates. CN-Cbl has absorption peaks in the 420 nm and 500-550 nm regions, which overlap with the blue and green light spectra used in this study. These results provide solid evidence supporting the view that different colors of light have different effects on CN-Cbl conversion and that only wavelengths above <600 nm cause CN-Cbl conversion.
Table 5. Colored light affects CN-Cbl conversion.

Use red light in manufacturing environments as a pink color control strategy. It was tested in a laboratory setting using red light instead of regular white fluorescent light. The reduced mAb A (Batch 2 product in Table 2) was spiked into the culture medium at the time of harvest at a mAb A concentration similar to the actual cell culture. The culture medium was previously exposed to white, red or green light (1.2 million hours). After incubation overnight (20 hours) in the dark, mAbs were purified on a protein A column and bound vitamin _B12 was measured in an ELISA assay. As shown in Table 6, vitamin _B12 deposition in media exposed to both white and green light was greater than those exposed to red light and dark controls. These results indicate that red light can indeed protect CN-Cbl from conversion.
Table 6. mAb A drug substance spiked into CB exposed to various colored lights

Example 2
Using hydrogen peroxide to prevent antibody disulfide bond reduction during the manufacturing process

Protein stocks two IgG4 molecules (1 and 2) and two IgG1 molecules (3 and 4) were used in this study. The IgG4 molecule had a single point serine to proline mutation in the original IgG4 hinge region motif CPSC, which resembles the IgG1 interchain disulfide bond structure CPPC (Liu H et.al. MAbs. 2012; 4: 17-23, Aalberse RC et. al. Immunology 2002;105:9-19). All mAbs were produced in CHO cell culture, and cell viability ranged from 50 to 90%. Clarified bulk (CB) was usually produced by depth filtration, unless otherwise specified, in which case it was produced by centrifugation. mAb 1 cell cultures in some cases were treated with low pH and dextran sulfate prior to depth filtration. Downstream purification is achieved by Protein A chromatography, an additional polishing chromatography step and a final ultrafiltration/diafiltration step into a histidine-sugar-based buffer. Protein concentration was determined by absorbance at 280 nm (A280) with a Dropsense-96 spectrometer (Trinean NV, 9050 Gent, Belgium).

Free sulfhydryl content (thiol) assay The method is based on the literature (Arner ES et. al. Methods Enzymol 1999; 300:226-239) with slight modifications, and uses 2,2'-dinitro-5,5'-dithiobenzoic acid (DTNB). )It was used. Briefly, 10 mL of clarified bulk or drug substance (DS) samples were mixed with 15 mL of TE buffer (50 mM Tris-HCl, 20 mM EDTA, pH 7.6) in a 96-well plate. DTNB solution (5 mg/ml in ethanol; Sigma-Aldrich, D8130) was freshly mixed with 8 M guanidine-HCl in 0.2 M Tris (pH 8) in a 1:9 volume ratio before use. 100 mL of DTNB/guanidine solution was added to each well and mixed. Duplicate samples were tested. The spectral absorption at 412 nm was measured and the molar extinction coefficient value of 13,600 M ⁻¹ cm ⁻¹ was used to calculate the free thiol content.

Redox indicator assay 2,6-dichlorophenolindophenol (D2932, Spectrum Chemical MFG Corp, Gardena, Calif.) was dissolved in water to make a fresh 1% stock solution. The oxidized form of 2,6-dichlorophenolindophenol was blue at neutral pH and became colorless upon reduction. At acidic pH, the oxidized form of 2,6-dichlorophenolindophenol was purple in color and turned colorless upon reduction. To determine the DCPIP color range for reduced and non-reduced samples, 100% mAb 2 cell lysate was mixed with normal clarified bulk produced by centrifugation to obtain a series of 0, 3, 5 cell lysates. , 10, 20, 30, 40, 50, 60 and 70% dilutions were purified. The free thiol content of each dilution was immediately determined and plotted as a function of dilution (Figure 6D). DCPIP (10 mL) was added to a series of cell lysis dilutions (1 mL at room temperature) and visual inspection determined which dilutions showed a color change within seconds to minutes.

Non-reducing and reducing Caliper
The LabChip GXII Touch HT System (PerkinElmer) was used for mAb purity analysis under both reducing (R_Caliper) and non-reducing (NR_Caliper) conditions according to the manufacturer's recommendations. NR_Caliper is the most convenient high-throughput method for separating and quantifying antibody fragments including double chain 1 light chain (HHL), single chain 1 light chain (HL), heavy chain (HC) and light chain (LC). It was a tool. Size exclusion chromatography was used to monitor intact mAb and high molecular weight (HMW) species.

Mass spectrometry measurements of antibody oxidation of mAbs Samples of mAbs 2 and 3 were reduced with dithiothreitol, alkylated with iodoacetamide, and digested with trypsin. Tryptic digests were purified using a Waters ACQUITY UPLC system (Milford, MA U.S.A.) before being analyzed on a Thermo Scientific Orbitrap Q-EXACTIVE™ PLUS mass spectrometer (Bremen, Germany). Separated by chromatography. A Waters Acquity BEH C18 column (1.7 mm, 2.1 £150 mm) was used for the separation (at 40°C). Peptide (mobile phase A: 0.1% formic acid in water; mobile phase B: 0.1% formic acid in acetonitrile) at a flow rate of 0.2 mL/min using a linear gradient of 2% to 80% mobile phase B over 110 min. ) was eluted.

The Q Exactive Plus mass spectrometer was operated in data dependent mode to switch between MS and MS/MS acquisitions. Ions were generated using a sheath gas flow rate of 40, an auxiliary gas flow rate of 10, a spray voltage of 3 kV, a capillary temperature of 275° C., and an S-Lens RF level of 60. Resolution was set to 70,000 (AGC target 3e6) and 17,500 (AGC target 1e5) for survey scans and MS/MS events, respectively. A dynamic exclusion time of 10 s was used with a single repetition number. Relative quantification was achieved by dividing the peak area of oxidized peptides by the sum of native and oxidized peptides in selected ion chromatograms of the Survey Scan.

Hydrogen Peroxide and Other Peroxide Treatments 30% stock hydrogen peroxide (Sigma-Aldrich cat. 216763, density 1.11 g/mL) was made into a 3% stock (v/v, equal to 980 mM) with water before use. Freshly diluted. 3% stocks were added to cell culture or CB to final concentrations of 0.1 to 20 mM as needed. To test the effect of hydrogen peroxide on antibody disulfide reduction, hydrogen peroxide stocks were added to the CB at final concentrations of 0, 0.1, 0.33, 1, 3, 5, 10, or 20 mM.

Two inorganic chemicals were also tested, including hydrogen peroxide, sodium percarbonate (Sigma-Aldrich, Cat. 371432, 2Na2CO3 3H2O2) and sodium perborate (Sigma-Aldrich, Cat. 372862, NaBO3 H2O). Both sodium percarbonate and sodium perborate undergo hydrolysis upon contact with water to produce hydrogen peroxide and carbonate or borate salts, respectively.

Small scale model of disulfide bond reduction A 500 mL bottle (FlexBioSys, Cat FXB-500M-0005) with a suction cap was filled with 300 mL clarified bulk. Alternatively, a 50 mL bioprocess bag (ThermoFisher, Cat SH30658.11) was filled with 30 mL clarified bulk. The vessel was connected to a pure nitrogen gas supply line and flushed with nitrogen gas (≦5 psi) for 1-2 hours at room temperature, then sealed airless for 1 day. After retention, samples were analyzed immediately in the free thiol assay or mixed with iodoacetamide (20 mM final concentration) before analysis on the NR_Caliper.

To generate the worst case of antibody disulfide bond reduction, 100% cell disruption was artificially induced by mechanical disruption. The cell culture was centrifuged (500 g, 20 min) and the cell pellet was dissolved in RIPA buffer (ThermoFisher, Cat 8990.25mM Tris, 150mM NaCl, 0.1% SDS, 1% sodium deoxycholate) at 1/10 of the original volume. , 1% NP-40, pH 7.6), shaken for 30 minutes at room temperature to dissolve, then cooled on ice and subjected to a high-speed tissue homogenizer for 3 minutes. It was assumed that complete cell lysis was achieved with these steps. The lysed material was then returned to the centrifuged supernatant. The cell lysate was filled into 50 mL bags with or without hydrogen peroxide, then flushed with nitrogen gas and kept airless at room temperature for 1 day as described above. After incubation, cell lysates were centrifuged twice at 3,000 g for 60 min at 4°C, passed through a 0.2 micron filter, and analyzed as described above.

Hydrogen peroxide treatment and its impact on product quality attributes For mAb 4, cell cultures in a laboratory scale bioreactor were used for the following two experiments. In initial experiments, clarified bulk was produced by centrifugation and filtration using a 0.2 micron filter. The resulting clarified bulks were either (a) stored at 4°C for 4 days as a control or (b) combined with hydrogen peroxide to a final concentration of 10 mM peroxide and stored at 4°C for 4 days. In the second experiment, hydrogen peroxide was added to the cell culture medium (containing cells) and incubated for 1 hour at room temperature before centrifugation. The resulting clarified bulks were either (a) stored at 4° C. for 4 days as a control, or (b) placed at room temperature on the third day, flushed with nitrogen gas, and kept at room temperature for another day.

All samples obtained were processed with protein A chromatography followed by a low pH virus inactivation and neutralization step using process center point conditions. Product quality was characterized by size-exclusion ultraperformance liquid chromatography to analyze intact antibody and aggregate levels, by capillary isoelectric focusing to image charge variation distributions, and under reducing and non-reducing conditions for purity. Evaluation was made using Caliper. Process impurities were estimated by ELISA for host cell proteins and qPCR for residual DNA.

For mAb 2, hydrogen peroxide was added to the clarified bulk at final concentrations of 0, 3, 5, and 10 mM and then kept at room temperature for 1 day under nitrogen gas flow. The resulting samples were processed via high-throughput Protein A chromatography prior to analysis.

Disulfide Bond Reduction During Manufacturing Reduction of mAbs was observed in process development for several IgG1 and IgG4 antibody molecules, either in clarified bulk (CB) or protein A elution (PAE).

Similar to other groups, we observed an accumulation of total free thiols in the clarified bulk during collection and primary recovery (WO2015/085003, Ruaudel J, et.al. BMC Proceedings 2015;9(9):24 ). The difference in free thiol concentration between samples of mAb 1 clarified bulk that exhibited disulfide reduction and those that did not was significant (FIG. 5A). For mAb 1, free thiol levels below 100mM in the clarified bulk had a low risk of low molecular weight production; free thiol levels between 100-200mM posed a potential risk; and free thiols above 200mM levels were at high risk of low molecular weight production. It should be noted that Ellman's reagent (2,2'-dinitro-5,5'-dithiobenzoic acid; DTNB), in addition to free sulfhydryl groups in proteins, can also react with small sulfhydryl-containing molecules such as reduced glutathione, cysteine, and lipoic acid. is to react with both sulfhydryl groups. Therefore, the baseline level of free thiol levels depends on the cell line, medium composition (small sulfhydryl-containing molecules) and cell viability. Nevertheless, when disulfide bond reduction occurred, free thiol levels tended to be much higher than baseline.

The level of free thiols in the clarified bulk is highly dynamic and is related to composition and processing process. In addition to air exposure, other factors such as storage temperature and pH also affect its dynamics. Adjusting the pH of the mAb 1 clarified bulk from 7 to 4.8 and then removing cells by depth filtration slightly decreased the free thiol levels and the observed percentage of low molecular weight (Figures 5B and 5C). When cell collections were treated with a combination of pH 4.8 and dextran sulfate (0.05 g/L and 60 min of mixing followed by centrifugation and filtration), free thiol levels were significantly reduced and flocculation was reduced by Trx/ Consistent with the observation that host cell proteins including TrxR and Glu/GR may be removed and some small sulfhydryl-containing molecules removed. The level of free thiols also varied with temperature. Incubation of the control clarified bulk at 37°C for 75 minutes caused a 60% reduction in free thiol content. However, when 1 mM NADPH, a necessary component for TrxR and GR reactions, was added to the solution before 37 °C incubation, an increase in free thiol levels was seen in control and pH-adjusted CBs (Fig. 5A). These dynamic changes indicate that free thiol levels should be monitored from pre-collection to post-collection steps to better predict the risk of low molecular weight production.

Using redox indicators to predict disulfide bond reduction Since disulfide bond reduction is a redox reaction, we use redox indicators as a simple, rapid and robust alternative to the DTNB test, and during collection and recovery. The potential risk of low molecular weight formation can be anticipated. Redox indicators undergo a distinct color change at a particular electrode potential in a similar way that a pH indicator undergoes a color change at a particular pH. In order to find suitable redox indicators, a series of commercially available redox indicators were tested and several were identified as the best potential candidates for use in the biopharmaceutical manufacturing process.

An example of a redox indicator is 2,6-dichlorophenolindophenol (DCPIP, Figure 6A). The color changes in mAb 2 purified drug substance with a high proportion of low molecular weight (Figure 6B), and the color changes completely and partially in the clarified bulk for high, medium, and low levels of free thiols, respectively. or no change (Fig. 6C). To determine the DCPIP color change range, DCPIP stock solutions were added to serial dilutions of mAb 2 cell lysates of known free thiol concentrations (Figure 6D). DCPIP underwent a color change in mAb 2 CB at 3-5% cell lysates and free thiol concentrations ranging from 80-100 mM. Therefore, if the clarified bulk sample had a free thiol concentration higher than 100 mM, DCPIP changes color, and this feature makes DCPIP a quick and easy-to-use method to predict the risk of disulfide bond reduction during manufacturing. .

ＮＡ：アプライせず A study was designed to investigate whether the color change in DCPIP correlates with free thiol levels and whether low molecular weight production could be correctly predicted (Fig. 6E). mAb 3 clarified bulk was generated by depth filtration with air or nitrogen gas flushing of the filter train before processing. Nitrogen gas flushing was used to create an "airless" condition where no oxygen was exposed to the liquid during the collection process. The resulting clarified bulk was aliquoted into small containers at 0, 20, 50 and 100% air overlay headspace volume to liquid volume and incubated for 1 day at 4°C, 24°C and 37°C. As shown in Table 7, the presence of air during collection and recovery was important in preventing low molecular weight production. In this case, the starting clarified bulk has a lower free thiol concentration (51 mM) compared to the airless condition (207 mM), and the dissolved oxygen (DO) level is lower even at % air overlay and holding conditions at different temperatures. It was sufficient (≧37%) to suppress low molecular weight production.
Table 7. Free thiol concentration and DCPIP color change in mAb 3 clarified bulk produced and maintained in various air and airless conditions.

NA: Do not apply

In contrast, % air overlay during retention became important for clarified bulk produced in airless conditions. The worst case was airless collection and airless retention, with intact mAb monomer reduced from 96.5% to 81.6%. Additionally, no disulfide reduction occurred when the air overlay was 20% or more. Incubation temperature had different effects on free thiol concentration. The highest free thiol concentration and greatest reduction occurred at the 24°C hold. The amount of free thiol after holding at 4°C was close to the amount before holding. Holding at 37°C was estimated to be the optimal temperature for TrxR and GR enzyme activity. However, all conditions universally decreased the amount of free thiols, and the data were consistent with the results for mAb1, as shown in Figure 5.

At the end of the one day retention test, each sample was mixed with DCPIP stock solution. Color changes were evaluated visually (Figure 6F). Table 7 shows that the color change sequence has a strong correlation with free thiol concentration; the higher the free thiol concentration, the faster the color change of DCPIP. All samples collected under aerated conditions showed no color change.

Additional redox indicators of different standard reduction potentials were tested in a similar manner and thionine and methylene blue were found to change color at free thiol levels of 600-800 and >1,000 mM, respectively (data not shown). do not have). These dyes can be assembled to include a ladder of indicators to quickly predict the level of risk during manufacturing.

A scale-down model to reproduce disulfide bond reduction during collection and primary recovery. Numerous lines of evidence suggest that disulfide bond reduction is due to cell lysis during the collection procedure and to insufficient air intake at room temperature over time. We suggest that this correlates with clarified bulk being held in closed containers (e.g. disposable bags).Mun M et. al. Biotechnol Bioeng 2015;112:734-742). Therefore, we mimic this phenomenon by filling a sealable container with a small amount of clarified bulk (10 to 300 mL) and flushing with nitrogen gas to maintain room temperature and then create an airless condition by NR_Caliper analysis for low molecular weights. We developed a small-scale model to do this (Figure 7).

Because clarified bulk from manufacturing operations can vary from lot to lot in terms of cell damage and holding times for various steps from pre- to post-harvest, the low molecular weight content can vary over a wide range from a few percent to nearly 100%. It can change. To simulate small-scale studies, cell lysates from CHO cells were mixed with the original volume of cell culture supernatant to simulate 100% cell lysis, and even in the presence of air, low molecular weight A worst-case scenario was generated (described in Materials and Methods above) (FIG. 8A), filled into a sealable container, flushed with nitrogen gas and incubated at room temperature for 1 day. Under these conditions, all intact mAb 2 antibody molecules in the clarified bulk were completely fragmented (Figures 8B and 9B). Free thiol content in 100% cell lysed clarified bulk was measured from 1 to 3 mM, which was at least 10 times higher than that of normal operation.

Hydrogen peroxide can inhibit the reduction of disulfide bonds.
One way to control the reduction of disulfide bonds in mAbs is to remove the reduction potential contained in the clarified bulk before the mAb is reduced. The use of H2O2 was tested for this purpose. As shown in Figure 7, the control mAb 2 clarified bulk produced approximately 20% lower molecular weight after being kept airless at room temperature for 1 day (Figure 7C). Hydrogen peroxide was added to the same clarified bulk at concentrations of 0.33, 1 and 3 mM prior to airless hold. The results showed that 3mM (approximately 0.01%) hydrogen peroxide could completely prevent the formation of low molecular weight (Figures 7E, 7G and 7I). These results indicate that hydrogen peroxide effectively prevents disulfide bond reduction.

The 100% lysed cell culture model described above was used to determine the minimum concentration of H ₂ O ₂ required to inhibit low molecular weight production in the worst case. In this study, 100% lysed cell cultures were maintained in airless conditions in the presence of 0, 5 or 10mM H2O2. After 1 day of holding, no intact mAb2 antibody molecules remain and low molecular weight species are predominated with a small fraction of halfmer light and heavy chains, compared to air-exposed controls. It shows that the reduction is almost complete compared to the sample (Figures 9A and 9B). Addition of 5mM and 10mM of H2O2 can partially (Fig. 8C) and completely (Fig. 8D) inhibit low molecular weight production, respectively. These NR_Caliper results from unpurified CB are summarized in Figure 8E. After protein A purification, 0, 5 and 10 mM H2O2 samples had antibody monomer purity of 0, 10.1 and 98.3%, respectively, from NR_caliper results (Figure 8F). The air exposed sample had 83% purity of antibody monomer. These results confirm that 10 mM H2O2 can effectively prevent mAb reduction in the worst case of clarified bulk processing.

There are many inorganic or organic peroxide-containing compounds that can replace hydrogen peroxide. Two inorganic forms, sodium percarbonate and sodium perborate, were tested as replacements for hydrogen peroxide. The results show that both peroxides can effectively suppress the reduction of disulfide bonds in a concentration-dependent manner (Figure 9).

表９．過酸化水素処理後のｍＡｂ ３メチオニン残基の酸化

Hydrogen Peroxide Treatment and Oxidation of Antibody Molecules One of the major concerns using hydrogen peroxide is the potential to oxidize mAbs. Methionine residues are most susceptible to hydrogen peroxide oxidation (Luo Q, et.al. J Biol Chem. 2011;286:25134-25144, Stracke J, et.al. MAbs. 2014;6:1229-1242 ). mAbs 2 and 3 (belonging to the IgG4 and IgG1 subclasses, respectively) were analyzed by mass spectrometry for methionine oxidation after various hydrogen peroxide treatments. For both molecules, after addition of H2O2 (from 1mM to 10mM) and incubation at room temperature for 1 day, a slight increase in oxidation depending on H2O2 concentration was observed in the same lot of samples under hydrogen peroxide treatment as described above. (Tables 8 and 9). However, the H2O2-induced oxidation changes for both mAbs 2 and 3 were small due to 1% variation in tryptic peptide mapping and LC-MS/MS methods.
Table 8. Oxidation of mAb 2-methionine residues after hydrogen peroxide treatment

Table 9. Oxidation of mAb 3-methionine residues after hydrogen peroxide treatment

These results indicate that under optimal peroxide concentrations, mAb oxidation mediated by hydrogen peroxide or other peroxides does not significantly increase, indicating that oxidation levels can be controlled.

＊ｍＡｂ ４の場合、処理は収集前細胞培養液または清澄化バルク工程のいずれかで行われ、得られた清澄化バルクは、プロセス中心点条件を使用してプロテインＡクロマトグラフィーおよび低ｐＨウイルス不活化および中和（ＰＡＶＩＢ）ステップをした。解析結果はＰＡＶＩＢ試料から得られた。ｍＡｂ ２の場合、過酸化水素処理およびコントロール清澄化バルク試料は解析前にハイスループットプロテインＡカラムにより精製した。ＮＤは「未決定」である。 Impact of hydrogen peroxide treatment on product quality attributes In addition to oxidation, the potential impact of hydrogen peroxide on other drug substance quality attributes and downstream purification processes was also evaluated in a laboratory-scale study. The experimental design included both mAb 2 (IgG4) and mAb 4 (IgG1). For mAb 4, hydrogen peroxide was added to pre-harvest cell culture medium and post-harvest clarified bulk (no cells). After various hydrogen peroxide treatments, the resulting clarified bulk was subjected to protein A chromatography followed by low pH virus inactivation and neutralization using center process point conditions prior to analysis. I did the process. The analysis results for mAb 4, shown in Table 10, showed no noticeable changes in all properties tested between hydrogen peroxide treated and untreated samples. Process impurities (host cell proteins and residual DNA) were also comparable to controls and tolerated over the process range (data not shown). Similar results were also observed for mAb 2 treated _with 0, 3, 5 and 10 mM _H2O2 at room temperature for 1 day (Table 10).
Table 10. Hydrogen peroxide treatment on product quality characteristics of mAbs 2 and 4.

*For mAb 4, processing is done either in pre-harvest cell culture or in a clarified bulk step, and the resulting clarified bulk is subjected to protein A chromatography and low pH virus deficiency using process center point conditions. An activation and neutralization (PAVIB) step was performed. Analytical results were obtained from PAVIB samples. For mAb 2, hydrogen peroxide treated and control clarified bulk samples were purified on a high-throughput Protein A column prior to analysis. ND is "undetermined".

Therefore, these results indicate that hydrogen peroxide treatment (up to 10 mM) has no noticeable changes in all properties tested in samples treated with hydrogen peroxide before and after cell removal.

Claims (8)

a) adding hydrogen peroxide to the clarified bulk at the time of collection at a concentration greater than zero but not exceeding 10 mM to prevent reduction of antibody disulfide bonds; and b) media preparation, media storage, and antibody production. and inhibiting the conversion of cyanocobalamin (CN-Cbl) to hydroxocobalamin (HO-Cbl) in the culture medium by preventing cell culture medium exposure to visible light during one or more selected times of antibody collection. A method of inhibiting vitamin _B12 binding to antibodies during cell culture production, comprising: 2. The method of claim 1, wherein the CN-Cbl is exposed to only red light (>600 nm) during media preparation, media storage, and optionally during antibody production and antibody collection. below:
a) culturing cells that produce the antibody of interest;
b) harvesting the antibody of interest from a cell culture; and c) adding hydrogen peroxide to the clarified bulk at the time of harvest at a concentration greater than zero but not greater than 10 mM. Culture method . A method of preventing antibody disulfide bond reduction comprising adding peroxide to the clarified bulk during collection at a concentration greater than zero but not greater than 10 mM. 5. The method of claim 4, wherein the peroxide is hydrogen peroxide. 5. A method according to claim 4, wherein the peroxide is an inorganic or organic peroxide. 5. The method of claim 4, wherein the peroxide is sodium percarbonate or sodium perborate. 5. The method according to any one of claims 1, 3 and 4, wherein the antibodies are selected from monoclonal antibodies (mAbs), human antibodies (HuMAbs), humanized antibodies and chimeric antibodies.

Images